Computer-Implemented Medical Analytics Method and System Employing a Modified Mini-Max Procedure

2. The method of claim 1 , wherein said mini-max procedure further comprises: 7) selecting by the computer one of said clinical data present (j); 8) selecting by the computer a clinical datum absent (r) in said patient that corresponds to diagnosis (k); 9) creating by the computer a clinical data pair (j, r) consisting of said one of said clinical data present (j) and said one clinical datum absent (r); and 10) calculating by the computer a partial probability P(k) j,r of diagnosis (k) for said clinical data pair (j, r) where the partial probability P(k) j,r is calculated using: ⁢ partial ⁢ ⁢ P ⁡ ( k ) j , r = PP ⁡ ( k ) j ⁢ ( 1 - S ⁡ ( k ) r ) PP ⁢ ( 1 ) j ⁢ ( 1 - S ⁡ ( 1 ) r ) + … ⁢ ⁢ PP ⁢ ( k ) j ⁢ ( 1 - S ⁡ ( k ) r ) + … ⁢ ⁢ PP ⁢ ( n ) j ⁢ ( 1 - S ⁡ ( n ) r ) . where step 7 follows step 6 and steps 7 through 10 are performed in numerical order.

3. The method of claim 2 , further comprising: 11) repeating the steps g) through j) to create additional clinical data pairs (j, r); and 12) calculating by the computer partial probabilities P(k) j,r for said additional clinical data pairs (j, r) such that said partial probabilities P(l) j,r , . . . P(k) j,r , . . . , P(n) j,r for corresponding ruled in diagnoses (k), where k =1 . . . n, satisfy a normalization condition: ⁢ ∑ k = 1 n ⁢ ⁢ P ⁡ ( K ) j , r = 1 where step 11 follows step 10 and steps 11 and 12 are performed in numerical order.

4. The method of claim 3 , wherein said mini-max procedure further comprises creating by the computer a mini-max table for each of said diagnoses (k) in said differential diagnosis list, whereby a first data column of each said mini-max table comprises said positive predictive values PP(k) j of said clinical data present (j) in said patient for said diagnosis (k) for which said mini-max table is created and a first row of each said mini-max table comprises said sensitivities S(k), of said clinical data absent (r) for said diagnosis (k) for which said mini-max table is created.

5. The method of claim 4 , further comprising the steps of: 13) transferring by the computer each said partial probability P(k) j,r into cells of said mini-max table where said PP(k) j for each said clinical datum present (j) and said sensitivity S(k), for each said clinical datum absent (r) converge; and 14) selecting by the computer from among said partial probabilities P(k) j,r in said cells a determining partial probability DP(k) j,r having the smallest value in its row and the greatest value in its column where step 13 follows step 12 and steps 13 and 14 are performed in numerical order.

6. The method of claim 5 , wherein from each said mini-max table, said determining by the computer partial probability DP(k) j,r is selected by the computer as a total probability TP(k) for said diagnosis (k) for which said mini-max table is created.

7. The method of claim 6 , further comprising confirming by the computer as final diagnoses (k) those of said diagnoses (k) for which said total probability TP(k) is greater than a confirmation threshold value CT, and ruling out by the computer those of said diagnoses (k) for which said total probability TP(k) is smaller than a deletion threshold DT.

8. The method of claim 7 , further comprising repeating by the computer the steps of claim 1 through claim 7 until said diagnoses (k) in said differential diagnosis list have satisfied said confirmation threshold CT and said deletion threshold DT.

10. The method of claim 9 , further comprising calculating by the computer a total probability TP(k) for each of said diagnoses (k) ruled in into said differential diagnosis list by including in said mini-max procedure said clinical datum not yet investigated (y).

11. The method of claim 10 , wherein said clinical datum not yet investigated (y) is treated by the computer as if present in said patient when computing said total probability TP(k) and wherein said clinical datum not yet investigated (y) is treated by the computer as if absent in said patient when computing said total probability TP(k).

12. The method of claim 11 , further comprising selecting by the computer said clinical datum not yet investigated (y) such that said cost C(y) is minimized and a change to said total probability TP(k) is maximized.

13. The method of claim 10 , further comprising simultaneously by the computer recommending collecting from said patient a plurality of clinical data not yet investigated (y) that minimize said cost C(y) and maximize a change to said total probability TP(k).

14. The method of claim 13 , further comprising calculating by the computer a total probability TP(k) for each of said diagnoses (k) in said differential diagnosis list with said plurality of clinical data not yet investigated (y) treated as if present in said patient, and then computing said total probability TP(k) with said plurality of clinical data not yet investigated (y) treated as if absent in said patient.

15. The method of claim 1 , further comprising determining by the computer whether the absence of any clinical datum among said clinical data (m) will decrease a total probability TP(k) of said corresponding diagnosis (k), and aborting said mini-max procedure for said clinical datum (m) when its absence does not decrease said total probability TP(k).

16. The method of claim 1 , further comprising the steps of: a) distinguishing competitive diagnoses (k) from concurrent diagnoses (k); and b) applying said distinction to diagnoses (k) ruled in on said differential diagnosis list.

17. The method of claim 1 , further comprising diagnosing by the computer complex clinical presentations comprising the steps of: 17) creating complex clinical presentation models listing all potentially related diagnoses (k); and 18) transferring all said potentially related diagnoses (k) to said differential diagnosis list when one confirmed diagnosis matches any of said potentially related diagnoses where step 17 follows step 6 and steps 17 and 18 are performed in numerical order.

18. The method of claim 1 , further comprising checking by the computer for interactions between drugs having the potential of masking any of said clinical data (m) belonging to a primary disease (i), comprising the steps of: 19) flagging all said disease (i) models in said knowledge base that include at least one clinical datum (m) susceptible to being masked by said drugs; 20) listing for said disease (i) models all said drugs and concurrent diagnoses (k) having the potential of masking any of said clinical data (m) belonging to said primary disease (i); 21) confirming presence of any of said drugs in said patient; and 22) removing any clinical datum (m) masked by said drugs from consideration if at least one of said drugs is confirmed present where step 19 follows step 6 and steps 19 through 22 are performed in numerical order.

19. The method of claim 1 , further comprising checking by the computer for interactions between concurrent diseases (k) having the potential of masking any of said clinical data (m) belonging to a primary disease (i), comprising the steps of: 23) flagging all said disease (i) models in said knowledge base that include at least one clinical datum (m) susceptible to being masked by said concurrent diseases (i); 24) listing for said disease (i) models all said concurrent diseases (i) having the potential of masking any of said clinical data (m) belonging to said primary disease (i); 25) confirming presence of any of said concurrent diseases (i) in said patient; and 26) removing any clinical datum (m) masked from each corresponding disease (i) that contains said clinical datum (m) in its disease (i) model where step 23 follows step 6 and steps 23 through 26 are performed in numerical order.